At 12:13 AM 8/16/2004, you wrote:
>...... In a number of areas
>of this roll, it was somewhat translucent, while in other areas,
>quite opaque.

Hi Terry,

I don't know if you have the capability to switch over to using the
internal CCFL that is built into the Dyna-CIS unit or not, but I have also
run across several translucent rolls where that was the solution.

For translucent paper, I found that the only way I could get a good scan
was to switch from using the external light source shining through the
paper to using the built-in light source to scan reflectively by bouncing
the light off the paper.

I also have tried switching the different color channels (red, green or
blue) from the CIS sensor using the jumper on Richard's board. So far, in
all the cases I've run across, the green channel always seems the best to
use for either the transmissive or reflective scanning cases. Your mileage
may vary.

Currently, I'm set up to scan either way (transmissive or reflective). The
conversion software also has to have a bit flipped to decode the scans
differently since one scan is the negative of the other. My CIS to MIDI
decoder tests for this and switches automatically, but I think you have to
make the change manually if you are using Richard's decoder software.

If I can be of help, I can try to make the scans on my setup if you can't
reach a solution with yours. My scanner setup limitations are with the
physical size and mounting of the rolls, but if they mount with the same
(or reasonably similar) endcaps as regular piano rolls, I should be able to
make them work. Email me off-list if you want to send a roll for me to test.

As a side note, I'm currently reprocessing all the scans I have previously
made, using the most recent matrix recovery algorithm software. I'm
inserting the keywords into the MIDI files and I have a Perl script written
that can read the embedded MIDI-type1 text events and automatically enter
the data into the data base and build a webpage from that, automatically
listing (and linking to) all the rolls in a directory on the server. I
hope to post many of my own scans in a week or two at most. Because of US
copyright law, I'll only post for direct, public download on the internet,
those scans from rolls published prior to 1923. However, I'll be making
everything available privately to the members of the rollscanner community
for the purposes of the archiving project.

Regards,
Warren